The field of art conservation science and practice relies increasingly on a multidisciplinary research including new sensing technologies and techniques for analyzing a wealth of available data. Multimodal imaging is now routinely employed during the restoration of paintings in order to detect more reliably regions or patterns of interest and to support thereby certain decisions that need to be made during the conservation-restoration treatments. In this talk, we discuss recent advances in digital signal processing and machine learning for supporting the actual restoration of paintings. The focus will be on sparse coding, representation learning, spatial context modelling with Markov Random Fields and Bayesian inference in tasks such as crack detection, paint loss detection and virtual inpainting. Concrete examples will be shown from the ongoing conservation-restoration treatment of the Ghent Altarpiece.
The presentation includes joint works with Ingrid Daubechies (Duke University), Maximiliaan Martens (Ghent University, Faculty of Arts and Philosophy), Bart Devolder (Princeton University Art Museum), Hélène Dubois (Royal Institute for Cultural Heritage, KIK -IRPA), Bruno Cornelis (Vrije Universiteit Brussel, Electronics and Informatics – VUB -ETRO) and Ann Dooms (Vrije Universiteit Brussel, Digital Mathematics group).